8 AI Tips for Web Developers (and Their Careers)

Artificial intelligence (AI) is sweeping the world, and web development is at the center of this AI revolution. What does this mean for our work?

Many developers may not have actively embraced AI tools yet. They may not have used AI widely to this day due to concerns about potential AI problems and their emphasis on their own professional skills.

However, it is a mistake to ignore the application of AI in web development. This article will share some experiences and techniques, which come from the author's practice of using AI in full in recent months.

(The following suggestions apply to individual developers. I will share my views in a subsequent article about the application of AI in the web development team. This article assumes that you have your own project, including your own website. )

1. Start Action

Get started using AI now. If you haven't used it yet, start today. Otherwise, your company may abandon you.

This is not a joke (or, I'm not optimistic).

AI is harmful to the environment and may also have an impact on our professional skills, but we cannot ignore reality. (Of course, we can and should take steps to address these issues, but this needs to be discussed separately.)

In recent years, layoffs in the technology industry have become increasingly serious. No matter how a company claims to be the "best employer", action is better than words. In a technology company, we must plan the worst.

2. Choose the right AI assistant

The AI ​​mentioned here refers to development-related AI. While general-purpose AI such as ChatGPT can also help, our work requires professional tools. GitHub Copilot is one of the most popular choices at the moment, and JetBrains' AI assistant is also a good choice. Get a license, try the tool that works for you, and most importantly, get started with it.

3. Automatically generate submission information

AI can automatically generate submission information, which can save a lot of time. (I have to process about 1000 submissions per month.) JetBrains' AI assistant can easily do this in the JetBrains IDE. Here is the prompt I used, which is based on the default prompt, uses the Conventional Commits specification and marks each commit as an AI generation:

• Avoid lengthy descriptions or unnecessary details.
• Start with an imperative sentence of no more than 50 characters.
• Use Conventional Commits style.
• Then empty the line and continue with a more detailed explanation.
• The first part only writes one sentence, and explains at most two to three sentences in detail.
• For anything in the repository..., just write the commit message as: "chore: include automated changes"
• Add the following message to the last individual line of each submission: (This commit message was AI-generated.)

It works great, saves time and is as good (or even better) as the information I wrote manually.

However, the following points are also important: Writing a standardized submission is still an important skill. If you are a junior developer and have not yet developed good code documentation habits, it is recommended that you develop this skill first and then use AI assistance. Mastering the basics is crucial because it is related to your professional credibility.

4. Use AI for code review

You may not be able to get code review unless you work for someone else. However, if you develop code in one organization, you may encounter various code review styles and processes – some reviewers approve everything, while others discuss tabs and spaces in each file ;In some cases you can get feedback immediately, while in some cases you need the support of managers to eventually get review from other teams.

In both cases, AI code review can help. If you can't get a second opinion right now, they are very useful and can speed up and coordinate code reviews within your organization.

What tools can you use? I'm still evaluating more options, but my favorite code review helper at the moment is CodeRabbit. It correctly identifies issues that I might not have noticed and I like to rely on it on more complex changes in my project.

Please note that I am not saying just relies on AI code review. While we may need to do this for our personal work, it is different in the organization. Here we can incorporate AI into the process and then adjust it based on our expectations and experience.

5. Refactor the code using AI

Like other developers (or like me), do you use to-do notes to mark what you can clean up? (or do you do that?)

A good use case for AI assistants or editors is to refactor these places. For example, JetBrains has an excellent refactoring option (I think this is its best AI feature besides committing information) that allows refactoring suggestions for such code. Try it on your own code. While the solution may not be "perfect", it may still be better than before. (If not, you can always add another "to-do" comment to your code - or roll back.)

Refactoring is where AI can come into play - I suggest you use it for this.

6. Complete five test items

Start now: Write down your biggest pain points at the moment, the most annoying repetitive tasks, and the projects you've always wanted to do.

If you have time and are already set up, have your AI code assistant provide scaffolding or scripting for the solution you need or the project you envision. Immerse yourself in it. Build it.

Complete five or more of these projects. If you are able to handle additional work (see "AI Paradox"), record, publish, and promote them.

The purpose is to develop the skills in AI development efforts and hone your perception of where AI is helpful (and where it is not).

7. Explore AI tools regularly

Continue to try AI solutions professionally and personally. Yes, everyone is releasing AI products, but the focus here is to focus on AI solutions in your field and to get into the habit of testing them (for me, repeated reminders are helpful for this).

The following are some solutions that I have been evaluating or planning to evaluate:

• Amazon Q Developer
• Bito
• Codacy
• Codepeer
• CodeScene
• CodiumAI
• Cursor
• Grit
• Figstack
• Mutable.ai
• PullRequest
• Reviewable
• Sourcegraph Cody
• Tabnine

After this is done, you and I can continue to explore many "excellent lists" for AI information. For example, Mahsima Dastan’s Awesome AI Tools or James Murdza’s Awesome AI-Powered Developer Tools.

8. Make sure you won't where you use AI

I have suggested waiting until you have developed a good habit of writing submission information before using AI to generate submission information, which most people will do.

This is part of a bigger topic – realizing what work you can delegate to AI and what might not be delegated.

I think this part is more subtle:

You can draw a line saying that you don't use AI to do the job you do best. For example, I don't use AI to generate HTML and CSS, and I have many reasons to do it myself (if you haven't done this yet, please consider following my work!).

But it still makes sense to test AI and check where it can improve your efficiency—or make you obsolete.

This looks like bad news, of course – but you want to learn this way before someone else tells you. This will enable you to adapt and adapt to a) integrate AI into your own work, and b) develop your non-AI image, identify and shape one that can benefit from you and your talents and experiences without having to do so from AI areas that benefit.

If you learn to use AI so that you know where AI can’t be used, then you’ll be in a place where we need professionals in the future. If you don't, the work you are doing will most likely be replaced and automated by AI.

Summary

1. AI is a large energy consumer, and we need to urgently find out how to reduce its resource consumption and use more renewable energy.
2. AI will make us stupid if we don’t think critically, continue to learn, and develop a perception of when not to use AI.
3. As developers, not only in web development, we need to incorporate AI into our work. There are enough companies that don’t care about people, so if we don’t adapt, we will be eliminated. (This is a disaster and needs change, yes.)
4. Integrating AI into our work starts with an AI IDE or an IDE with an AI code assistant.
5. AI generation of submission information is a good AI use case.
6. AI-supported code review is a mandatory AI use case. (I am absolutely optimistic about this.)
7. Refactoring supported by AI is a good AI use case.
8. We can accumulate experience and a keen perception of AI by using AI to solve pain points and implement small projects.
9. We benefit from focusing on the AI ​​field and regularly exploring AI tools.
10. We have to be very knowledgeable about AI so that we also know the shortcomings it relates to our work – and we have to use that knowledge to constantly redefine ourselves as professionals.
I believe this is the current situation and what we can do in web development now.

Thanks to Kevin Khaw for reviewing this article.

The above is the detailed content of 8 AI Tips for Web Developers (and Their Careers). For more information, please follow other related articles on the PHP Chinese website!